Early detection and intervention for children with ASD leads to better outcomes (e.g., Dawson et al., 2010). Intervention depends on accurate identification; therefore, early screening and diagnosis are crucial (Filipek et al., 1999). The Modified Checklist for Autism in Toddlers (M-CHAT) has become the most widely used early autism screening tool for children 16-30 mos. However, though the M-CHAT and other screening instruments have been available for 10 years or more, and universal screening has been endorsed by national organizations, the median age of ASD diagnosis in the U.S. is still four years or older. Children diagnosed at four years or later miss the possibility of crucial earl intervention. Furthermore, no screener can detect every child; therefore, physician surveillance of children's development, which involves integrating behavioral observation with screening, parent concern, and history information, is essential. Therefore, although a strong screening tool is necessary for effective screening, it is not sufficient on its own. Effective screening policies must consider issues such as when screening should occur, how best to implement screening in the context of ongoing developmental surveillance, and what child, family, physician, and practice factors influence the success of screening and surveillance. The proposed study builds on the seminal M-CHAT work, which did the foundational work for an effective screener. This study will investigate three key questions: (1) what is the best age to initiate screening (12, 15, or 18 mos)? (2) Can accuracy of physician surveillance be improved with a brief, focused intervention? (3) What are the factors associated with disparity in effective ASD screening and surveillance at the child/family level (e.g., race/ethnicity, socioeconomic status, sex) and at the physician/practice level (physician beliefs and self-efficacy, practice location, and patient demographics)? With a total of 12,000 toddlers screened for ASD, this study will examine how the variables in these three aims influence age and accuracy of ASD detection, allowing this team to identify targets for future intervention to reduce disparities and facilitate early and accurate diagnosis for all children. Much of the work done in early ASD screening comes from the team assembled for this proposed study, and the large sample of children from diverse geographic areas will make these results immediately generalizable. The proposed study will answer key remaining questions about early ASD screening, thereby impacting the early health monitoring of children across the nation, and helping ensure that the promise of early ASD detection can become a reality.